XuanShu API documentation

Find the right guide from tool setup to platform management

Find guides for Claude Code, Codex, Gemini CLI, the OpenAI SDK, compatible APIs, models, streaming connections, usage, and billing.

Three article groups · Clear paths · Practical answers

Client setup guides

Pick your client and go from install to a verifiable first request, with one-line scripts, manual setup, and troubleshooting.

Install Node.js

Install Node.js 22 or newer on Windows, macOS, and Linux.

Anyone preparing to run AI coding tools.

Resolve unavailable node and npm commands.

Connect Claude Code to XuanShu API

Set up Claude Code with a one-line script or manual configuration.

Developers and teams working with Claude Code.

Resolve Base URL, API key, model selection, and request verification questions.

Connect Codex to XuanShu API

Configure Codex CLI and Codex App to call currently available models.

Users who build software with Codex.

Resolve config.toml, authentication, and connectivity questions.

Connect Gemini CLI to XuanShu API

Configure Gemini CLI for the compatible XuanShu API entry point.

Individual developers and technical teams that prefer terminal workflows.

Resolve credential setup, model selection, and command-line verification questions.

Connect Hermes to XuanShu API

Select a channel and write the Hermes config with a one-line script.

Developers using the Hermes agent.

Resolve channel choice, config.yaml writing, and endpoint differences.

Connect OpenClaw to XuanShu API

Configure OpenClaw for the Anthropic or OpenAI channel.

Developers using OpenClaw.

Resolve installation, channel choice, and configuration questions.

Connect TRAE SOLO to XuanShu API

Add a custom model in TRAE SOLO.

Users of graphical clients.

Resolve custom endpoint and model name entry questions.

API / SDK

Choose a compatible API or SDK path to connect existing code to XuanShu API.

Anthropic Messages examples

Call the Messages API for text and image understanding.

Developers calling the Claude interface directly.

Resolve request shape, streaming, and image input format questions.

Image generation and editing examples

Use gpt-image-2 for text-to-image, image-to-image, and editing.

Developers who need image capabilities.

Resolve image parameters, streaming events, and result saving questions.

Responses API

Organize inputs and read outputs through the Responses-style interface.

Developers building newer OpenAI-style integrations.

Resolve input formatting, output parsing, and migration questions.

Chat Completions API

Send requests through the compatible Chat Completions interface.

Developers maintaining existing chat completion integrations.

Resolve request shape, streaming options, and response parsing questions.

Use the OpenAI SDK with XuanShu API

Reuse the OpenAI SDK while switching its endpoint to XuanShu API.

Developers with existing Python or JavaScript SDK projects.

Resolve client initialization, Base URL, API key, and model configuration questions.

Using the platform

Understand model selection, streaming connections, usage, billing, and capability boundaries.

View current models

Confirm currently available models through the model list.

Users selecting a model or checking a model identifier.

Resolve changing availability and uncertain model name questions.

Streaming WebSocket

Understand streaming WebSocket connections and event handling.

Developers building real-time interactions and lower-wait experiences.

Resolve connection setup, event consumption, and interruption handling questions.

Usage and billing

Understand credits, subscriptions, pay-as-you-go billing, and call records.

Individual developers, team administrators, and purchasing evaluators.

Resolve budget planning, usage reconciliation, and billing option questions.

Model capability matrix

Compare the supported scope of retained protocols and model capabilities.

Developers and architects evaluating a technical approach.

Resolve capability, streaming support, and limitation comparison questions.